Using this year's Deepwater Horizon oil spill in the Gulf of Mexico as a real-time experiment, scientists have developed a new model that can predict the movement of contaminants with greater accuracy than ever before. Their model gave several days of warning to authorities about where the gulf's currents would carry the huge swaths of BP oil, and the researchers say the model can predict how any hazardous or unwanted material will spread.
When fluid dynamicist Igor Mezić of the University of California, Santa Barbara, watched TV reports in late April showing 10,000 or more barrels of crude oil a day spilling into the gulf, he quickly assembled a research team. Mezić, who has been studying the mixing dynamics of fluids, such as oil and water, for 20 years, says he thought he could develop a better method of predicting where the oil would spread, thereby giving workers more time to deploy containment and cleanup equipment.
By early May, the researchers were trying to track the progress of the oil spill using existing fluid dynamics models and applying them to aerial and satellite images of the area of the gulf surrounding the spill site. But Mezić says they "were not happy" with the results. So his team incorporated recent equations he had developed into U.S. Naval Research Laboratory ocean-current data. The resulting "visualization software" can reproduce how an oil plume in a given body of water will spread.
The new model revealed that a fluid dispersing within a larger fluid—such as an oil plume in the ocean—tends to break into long, thin filaments instead of a single, gradually expanding cloud. Mezić calls them "stretching events," and they're caused by the chaotic attracting and repelling interaction of the water and oil, powered by the currents of the gulf, which flow at different rates and in different directions.
"That meant some of the oil could move very rapidly toward the coastlines of Louisiana, Mississippi, and Florida, while other parts of the spill would stay relatively far out to sea," says Mezić, whose team reports its findings online today in Science. The key factor, he explains, is that the water distributes the oil in a chaotic way, depending on average velocity of the currents. "We did not realize prior to this work just how to capture all of the aspects of mixing and stretching in such a complex flow."
Using the new formulas, the team was able to predict accurately, several days in advance, that the oil would first reach the shores of Plaquemines Parish and Grand Isle, Louisiana, and Pensacola, Florida, in late May, and then Panama City Beach, Florida, in early June. Mezić says that, based on the revelations of the model, his team members stayed in telephone contact with the spill cleanup team, telling them, days in advance, where the oil was heading and where they could best deploy their people and resources.
One of the biggest problems in the Deepwater Horizon spill was finding the dispersed oil slicks out at sea, where they could be contained and removed more easily than when they reach coastal waters and wetlands. But in future spill events, Mezić says, "our work should make this easier to do." In addition, he says the new model should help scientists better predict the expansion of other troublesome environmental events, such as volcanic ash clouds or anywhere a contaminant is being moved around by wind or water.
There is "reasonably good agreement between the geometric patterns produced by the model and the satellite/aircraft-derived patterns of the [surface oil] in the gulf," says mathematician Andrew Poje of the City University of New York, College of Staten Island. The model does so, he explains, by computing the "skeleton" of the particle dynamics—where the oil filaments form and what the geometry is when those filaments get stretched out by the water currents.
For more on the gulf oil spill, see our full coverage.